Optimization of data access may be performed using data access statistics. Data access statistics may include tracking writes and/or reads of a block, segment, or other portion of data. Data that is accessed more frequently may be considered as having a higher Input/Output (I/O) temperature. The data access statistics and/or I/O temperature of data may be used to make storage decisions about data (e.g., data with hotter I/O temperatures may be stored in storage with faster access times). However, data may be deduplicated, which may cause inaccurate data access statistics and/or loss of data access statistics. Specifically, duplicate segments or blocks of data may be eliminated and replaced with references to a single copy of the segments or blocks. Such deduplication may reduce storage needs, but eliminate and/or ignore data access statistics associated with the deleted duplicate segments or blocks. Such deduplication may result in incorrect data access statistics and/or I/O temperatures for a deduplicated block or segment.
In view of the foregoing, it may be understood that there may be significant problems and shortcomings associated with current data access statistics management technologies.